Research note: This material is for educational and strategy-design purposes. It is not investment advice or a promise of future returns.
Executive summary
Earnings-revision strategies buy stocks with improving analyst or management earnings expectations and avoid those with deteriorating expectations. The research base is closely related to earnings momentum and underreaction, and it is particularly useful in India around quarterly result seasons and management-guidance cycles.
This is an information-diffusion strategy. It assumes that the market does not fully incorporate revisions immediately, especially when the revision is credible, broad-based and supported by price confirmation. In practice, revision signals are stronger when combined with liquidity and price momentum.
Description
The cleanest version uses consensus EPS upgrades over the last one to three months. Where a full analyst database is unavailable, an India-specific proxy is management-guidance changes, sell-side target-price revisions from a trusted provider, or rolling upgrade frequency after corporate results. The theory is aligned with the literature showing drift after forecast revisions and earnings-related information releases.
Because India has concentrated analyst coverage in large and mid caps, the strategy should be universe-aware. Large caps may have faster information incorporation, but small caps can have patchier coverage and noisier signals. A balanced implementation often lives in Nifty 200 or a liquid large-mid universe.
Key attribute table
The range below is a practical research target inferred from revision and momentum literature; net returns depend heavily on data quality and coverage.
| Time horizon | Turnover | Typical Sharpe or return expectation | Data needs | Complexity | |---|---|---|---|---| | Short to medium term | Medium to high | Sharpe roughly 0.5–1.0; annualised return target roughly 10–18% gross | Analyst-estimate history, results dates, prices | High |
Details
Universe: Nifty 200 by default. Signal: percentage change in next-twelve-month EPS consensus over the last one month and three months, standardised by the historical volatility of revisions. Add confirmation from recent price momentum or abnormal-volume reaction to reduce false positives. Rebalance weekly or monthly around results seasons. Required inputs include exact result dates, revision timestamps and consensus-provider history; if unavailable, a reduced-form implementation can use management guidance, order-book disclosures or margin-upgrade commentary from structured text sources.
Risk controls should include a blackout window around same-day results if the desk cannot process revisions intraday, a minimum analyst-count threshold to avoid one-broker distortions, and sector balancing because upgrades often cluster by cycle. Position sizing can be linear in revision z-score, capped by liquidity and earnings-event risk.
Backtests must use timestamped revisions, not final historical consensus. Edge cases include companies with sparse coverage, abrupt corporate reshaping, one-off tax changes and split or bonus effects on EPS history. If you cannot source clean point-in-time estimates, do not pretend that the backtest is institutionally valid.
Implementation guide
- Source point-in-time consensus EPS data or a credible revision proxy.
- Align each observation to exact result and revision timestamps.
- Compute one- and three-month revision scores.
- Confirm with price action or abnormal volume.
- Build a diversified basket with liquidity and analyst-coverage filters.
- Rebalance around the weekly or monthly signal cycle.
India-specific example
Assume INFY, HAVELLS, TRENT and SUNPHARMA all reported results this quarter. If INFY’s next-twelve-month EPS consensus rises from ₹72 to ₹76 over one month, the revision is +5.6%. If HAVELLS falls from ₹56 to ₹53, the revision is -5.4%. All else equal, INFY belongs in the long candidate set and HAVELLS moves down or out. The actual production signal should use timestamped revisions from a licensed point-in-time source or a robust internal proxy built from result commentary and estimate updates.
Suppose a ₹40 crore revision sleeve wants 15 names. The strongest upgrade cohort could receive 8–9% of NAV collectively in the top three names, while weaker positive revisions receive smaller weights. A cautious PM would defer trading until after the initial event-day spread normalises in the regular session, especially because Indian result days can produce large open gaps followed by intraday reversal. Cash-market taxes and liquidity still matter because turnover is materially higher than in slow-moving value or quality sleeves.